1. Field of the Invention
The present invention relates to document OCR implementing devices and document OCR implementing methods, and more particularly, to a document OCR implementing device whereby a document is processed by individual OCR engines corresponding to recognition images, and a document OCR implementing method.
2. Description of the Related Art
Recently and continuingly, a network communication system having a document input and output device with a document OCR implementing device, the document input and output device being connected to a network and using plural communication protocols, the document input and output device communicating documents in various data forms to and from plural information devices, has been developed.
In such a network communication system, various application services wherein a document input and output device is used as a core are provided. For example, a document image read out or data made by an information device is sent to a designated address by e-mail, sent by a facsimile, or file-transferred to another information device. Information written in a received e-mail or an image in a file attached to the e-mail is recorded and output, sent to a designated facsimile, or file-transferred to the information device. Storing management of the data sent to the device is performed. See Japanese Laid-Open Patent Application Publication No. 2004-356822, for example.
In a case where an image input from the outside is digitized and stored in a storing device for common use in the network and the image is a subject of search, character recognition (OCR) processing of the document by the OCR implementing device is performed. Here, the image input from the outside is, for example, an image formed by reading a manuscript with an image reading device, an image received by facsimile, an image downloaded while reading a Web page, or the like. See Japanese Laid-Open Patent Application Publication No. 2004-356822.
In the document OCR implementing device having the above-mentioned structure, if a large amount of the documents are character recognition processed, the precision of recognition is improved. However, a large amount of time is required for the processing. In addition, it is not possible to obtain results of the character recognition processing corresponding to an object desirable for the user.